TAO/Triton is a little more conservative in their readings than CPC, and it still has a >+4c area in the western subsurface

I went through, since data became available in 1985, and found the 10 top analogs to broad subsurface state in January-February, like we have now. There are 5 positive analogs, 5 negative analogs. It should be noted that a lot of analogs had a warm/cold western part vs a cold/warm eastern part, in contrast, and I didn't use those, I only did the whole subsurface based warm or cold. 

Jan-Feb 2026 so far

-1) 2025 -- This later in the year was ENSO Neutral (0-1-0), with the ONI peaking at -0.5c, but not for 5 consecutive months. The RONI in 26-27 was Weak-La Nina, -1.0c

+1) 2023 -- This later was a Strong El Nino (1-1-0), ONI peaking at +2.1c, RONI peaking at +1.5c

+2) 2015 -- This later was a Strong El Nino (2-1-0), ONI peaking at +2.8c! RONI peaking at +2.4c

+3) 2014 -- This later was a Weak El Nino (3-1-0), ONI peaking at +0.8c, RONI peaking at +0.6c

-2) 2007 -- This later was a Strong La Nina (4-1-0), ONI peaking at -1.6c. RONI peaking at -1.6c

+4) 2002 -- This later was a Moderate El Nino (5-1-0), ONI peaking at +1.3c. RONI peaking at +1.5c

-3) 1999 -- This later was a Strong La Nina (6-1-0), ONI peaking at -1.7c. RONI peaking at -1.7c

+5) 1998 -- This later was a Strong El Nino (7-1-0), ONI peaking at +2.4c. RONI peaking at +2.4c

-4) 1994 -- This later was a Moderate El Nino (7-1-1), ONI peaking at +1.1c. RONI peaking at +1.4c

-5) 1988 -- This later was a Strong La Nina (8-1-1), ONI peaking at -1.8c. RONI peaking at -1.9c

So out of the hand-picked 10 best analogs of the past 40 years, 8/10 (80%) became the same ENSO state (warm subsurface to El Nino, cold subsurface to La Nina), 1/10 (10%) was Neutral the next year, and 1/10 (10%) was opposite. The total cumulative ONI peak for the 10 top analogs since 1985, was +13.9c, an average of +1.39c per year (Moderate ENSO state). 

CPC in January gave these probabilities

I now see the RONI is linked on the CPC page. Very good! (Edited analogs above -- the RONI is good because it neutralizes the global warming skew, where ENSO has warmed on average about 0.2-0.3c vs historical)

Climate Prediction Center - Relative Oceanic Niño Index (RONI)

Mid January IRI models forecast. I will edit these if someone links more recent forecasts. 

Since 1994, only 7/31 ENSO years have been Neutral, per RONI. 24/31 have either been La Nina or El Nino. 

Since 1994, only 7/31 ENSO years have been Neutral, per ONI. 24/31 have either been La Nina or El Nino. 

Per RONI, 5 of the last 6 years (20-26) have been La Nina.  Per RONI, 7 of the last 10 years (16-26) have been La Nina.  Per RONI, 11 of the last 19 years (07-26) have been La Nina.   Per RONI, 15 of the last 28 years (98-26) have been La Nina. That's >50% La Nina since 1998. 

You can see this post-1998 general La Nina pattern clearly in the Northern and Southern Hemisphere's Hadley Cell, High pressure right above and below Nino 3.4:

^People say it's a decadal PDO state, I think it's a decadal cold-ENSO state. That map encompasses the most recent 333 months. 

How has forecasting changed since then?  The AFDs are certainly different now with less analysis and more synoptic descriptions.  

I see the models now (HRRR, NBM, ensembles, AI) are significantly more accurate than the previous generations.  If a forecaster has a different opinion than the model consensus, what tools do you use to formulate your opinion?  Can a NWS meteorologist go against model guidance?  Is there occasionally a need to go against model guidance?

I get the feeling in the past there was more personality and subjectivity in operational forecasting than there is today.  Is this true?

For anyone new wanting to enter, the forecast contest asks you to predict temperature anomalies at nine locations across the U.S., in F deg relative to 1991-2020 averages.

The nine locations are DCA _ NYC _ BOS _ ORD _ ATL _ IAH _ DEN _ PHX _ SEA

... Happy new year ...

Based on ENSO and the teleconnections, there is a signal for the coldest month relative to normal to be January in the Gulf Coast and Southeast. Perhaps the Arctic outbreak of November 10-11 that saw the coldest anomalies affect the Southeast offers a hint of how things might unfold in January.

The C3S multi-model system appears to be quite aggressive with above normal temperatures. Last winter, there were some indications that it was too aggressive with the warmth, but I deferred to the guidance. That proved to be a bad mistake leading to an overly warm idea.

This time, I am giving more weight to the analog cases for insight. The ECMWF seasonal forecast appears to be a better fit with the final cases I selected based on ENSO and the teleconnections.  My selected comparison cases for Winter 2025-2026 are Winters 2016-2017, 2017-2018, 2021-2022, and 2024-2025. 

C3S Seasonal Forecasts:

ECMWF:

Analog Cases:

ECMWF: December-February 500 mb Progression:

Analog Cases: December-February 500 mb Progression:

Select Snowfall Estimates:
Albany: 45"-60"
Baltimore: 10"-15"
Binghamton: 80"-90"
Boston: 35"-45"
Buffalo: 85"-95"
Burlington: 75"-85"
Caribou: 110"-125"
Chicago: 30"-40"
Detroit: 40"-50"
New York City: 15"-25"
Newark: 15"-25"
Philadelphia: 10"-20"
St. Louis: 10"-15"
Toronto: 100 cm - 125 cm
Washington, DC: 8"-12"

______ DCA _ NYC _ BOS __ ORD _ ATL _ IAH ___ DEN _ PHX _ SEA 

... deadlines will be 06z of 1st of each month, and late penalties 1% per half day to 18z 4th, then 1% an hour. (7% + hrs, expiry 8th 15z)

Contest will continue regardless of participation rates even if it turns into a smaller group, as we strive for the elusive new record of 810/900.

A history of this contest (2013-2024) is posted in Dec 2023 contest and will migrate to Dec 2024 at end of contest year. 

Don't forget, no contest threads in 2025, all months will be contained and scored in this one thread. Good luck everyone, and welcome to any new participants. (scroll thru Dec 2024 contest to get an idea of what we do).

RONI vs. ONI.xlsx

...Enderlin Tornado #1 Upgraded to EF-5...

The National Weather Service in Grand Forks conducted additional surveys and worked extensively with wind damage experts to further investigate the 20 June 2025 Enderlin, ND tornado #1. The estimated maximum wind speed of the Enderlin, ND tornado #1 is greater than 210 mph and occurred during the time of the train derailment south of Enderlin, ND. The analysis involved forensic damage wind speed estimates for tipping several fully-loaded grain hopper cars and lofting of tanker cars, including one empty tanker car that was tossed about 475.5 ft (145 m). The maximum wind speed also correlates to the maximum strength on the WSR-88D Storm-Relative Velocity data from KMVX. 

Additional high-end damage indicators that have been reanalyzed and adjusted to include damage to the trees near the Maple River east of Enderlin and damage to farmstead #2 on Hwy 46. Further analysis of the trees surrounding the Maple River show extensive tree damage throughout the entire river valley with only stubs of large branches or trunks remaining and debarking with a "sandpapering" effect prevalent. Trees with attached root ball displacements were noted, including one where the original location could not be determined. Wind damage experts analyses helped determine the damage at farmstead #2 to be complete destruction with the foundation considered to be  swept clean, and debris swept downwind and scattered. However, issues surrounding the lack of proper anchoring will limit the overall rating in this area. Trees around the farmstead were also debarked, and saw root ball displacements. 

The National Weather Service in Grand Forks would like to thank Tim Marshall, Jim LaDue, Dr. Connell Miller, Dr. Greg Kopp, Dr. David Sills, and the entire Northern Tornadoes Project at Western University's Canadian Severe Storms Laboratory team for their assistance. 

Rating: EF-5

Estimated Peak Wind: Greater than 210 mph

Path Length /statute/: 12.10 miles

Path Width /maximum/: 1850 yards / 1.05 miles

Fatalities: 3

Injuries: 0

Start Date: 06/20/2025

Start Time: 11:02 PM CDT

Start Location: 3 S Enderlin / Ransom County / ND

Start Lat/Lon: 46.5728 / -97.6031

End Date: 06/20/2025

End Time: 11:21 PM CDT

End Location: 5 SSW Alice / Cass County / ND

End Lat/Lon: 46.987 / -97.5803

Full review, including DI images:

https://www.weather.gov/media/fgf/Enderlin.pdf

If there are any suggestions, let me know. 

Thanks!

https://www.weather.gov/media/notification/pdf_2025/pns25-37_mag_webpage.pdf

Site in question.

https://mag.ncep.noaa.gov/model-guidance-model-area.php

For instance at my location in Michigan, the low temperature for the 24 hour period from 12am to 12am defined as that official "day" usually occurs around 6am of that "day".  Is the "Low" forecasted temperature the temperature from 6am of the 12am to 12am period defined for that official day?  Or is the "Low" forecasted temperature actually the lowest temperature from the morning of the next official "day" typically around 6am ("overnight low??").   

I have been using Weather.com so I would like to know what their definitions are but also want to know how all the major reporting channels define these.

However, given the -PDO, collapse of the subsurface warmth and historical precedence following a nino this strong, I am favoring the development of La Niña in the May-June timeframe. If this turns out to be correct, I would expect the La Niña to strengthen from weak (-0.5 to -1 ONI) to strong (< -1.5 ONI) by fall. There is some guidance that keeps the La Niña as a weak event, but I am highly skeptical of that. 

First, where is the money and what is the "middle" of what people -- insurance companies, energy companies, etc., are predicting. 

Natural Gas Futures

Last year, the Natural Gas price in the Fall was extremely low, especially compared to Crude Oil/Gasoline, and this was a better predictor than seasonal models, which generally had a trough on the East Coast, US, from an expected El Nino pattern. 

Current Natural Gas price is $2.64. 

Here is a chart going back to 1998, highlighting the highest (blue) vs lowest (red) prices. 

Higher Natural Gas price should be linked to colder weather in the eastern US, and Europe. and Lower Natural Gas price should be linked to warmer weather in the eastern US, and Europe. 

Here are the Highest Natural Gas years (blue):

Here are the lower Natural Gas price years (red):

For October, here is where we rank:

1. 1998: $2.27

2. 2015: $2.3

3. 2019: $2.63

4. 2024: $2.64

5. 2023: $2.91

6. 1999: $2.95

26: 2006: $7.53

27: 2007: $8.33

28: 2004: $8.72

29: 2005: $12.2

Furthermore, because of inflation, the Crude Oil or Gasoline vs Natural Gas spread is a better gauge for relative value. 

We are currently #2 in this metric [since 1998], behind only last year (23-24). 

ENSO

I manually plotted all ENSO variables (200mb wind, 850mb wind, OLR, SSTs, SOI, pressure, MEI, etc.), and I found that the most correlated ENSO measurement to the North Pacific [PNA] pattern, is ENSO subsurface. This works at +0-time. 

Because of this, I use the subsurface primarily to determine what the ENSO state is, and is going to be for the Winter. Obviously, in the future, it could change, but right now we are completely Neutral. 

For most of the year so far, we have been in a "La Nina" in the subsurface: 

This has correlated with a -PNA pattern

Now that the subsurface has neutralized, the PNA is not correlating so highly

NAO

In 2005, I found that the N. Atlantic SSTs in a region from New Foundland to Greenland  from May-Sept has a high correlation to the following Winter's NAO. The correlation was almost 0.5 (or 75% of getting the sign right). I made a manual index of the region, and have followed its predictions every year since 2005. I estimated that this NAO predictor index has a 0.54SD at getting the Dec-March NAO correctly (+1.00 index is 50% odds of +0.46 to +1.54 DFJM NAO). In real time, that method has been 9-9 on the 0.54 SD since Inception, and it has gotten the phase correctly 13-5. That is real future time forecasting results. Here is what the index encompasses:

I weight the index as follows:

It's been working out great in real-time. 

This year, the index comes out at: 

Top area: ~0.0 (x1.00)

Bottom area: ~+0.8 (x0.65)

Total: +0.52

+0.52 NAO predictor for Dec-Jan-Feb-Mar. 

That means there is a 50% chance the DJFM NAO will be -0.02 to +1.06 (using 0.54 standard deviation)

That gives a 74% chance of the Winter DJFM NAO being Positive overall. 

PDO

I have been burned on the PDO! I do not think SSTs lead, I think they are more secondary to atmospheric conditions. But the last 4 years, and actually the last 30 years, the PDO has performed admirably. The mathematical odds are something like 1/100 for random to hit as much as the PDO has over this time. Because of that, I will give it some credence. 

CURRENT PDO IS NEAR -3. 

That is 2nd on record for October, going back to the early 1900s. Only 1955 had a lower October PDO.

Here is what PDO correlation looks like in the Winter (map default is the "+" phase, you have to flip it around to get a negative PDO correlation). 

Rolled-forward North American Temps

December 2023 to August 2024 was the warmest on record for the CONUS, due mostly to +EPO pattern. 

I made an analog list of 30 matching analogs (75 total years in dataset.. 30 analogs is 40%) and I got a really strong signal the following Nov-March. 

When you have 40% of the dataset used, you'd expect the anomalies to come out at +1F, but what I found was a much stronger signal than that:

Dec-June analogs:

Following Winter (40% of dataset!):

Mexican Heat Wave in May

Mexico crushed records in May. I found that similar analogs rolled above average temperatures to the eastern 1/2 of the CONUS for the following Sept-March. 

Phoenix Heat Wave

It's the warmest Sept 25 - Oct 13 in Phoenix all time. I think they broke their 2-week record by more than +7F!

I came up with 20 + analogs and found this for the following Winter rolled-forward:

Wintertime 10mb

QBO correlates with the Winter time 10mb state. When coupled with ENSO, its correlation is very strong. Going back to all records available, the correlation is 75% that +QBO/La Nina leads to a negative Wintertime 10mb in the Northern Hemisphere. 75% that -QBO/El Nino leads to a positive Wintertime 10mb in the Northern Hemisphere. 

The QBO is currently +10 and rising. It will likely peak in the Winter. >10 events for the QBO are a strong phase. With La Nina tendencies occurring, although I'm not necessarily predicting a La Nina, I think the odds favor a cold 10mb N. Hemisphere vortex by about 2/3 or 67%. 

Cold 10mb is correlated to +AO conditions in the Wintertime. 

There are other things I have considered, that I may talk about later but here is my Winter forecast:

Winter Forecast

Temps:

Precip:

I'd put my confidence as follows:

vs 10-year average: 65%

vs 30-year average: 75-80%

vs 50 year average: 80-85%

Obviously we have to wait for the survey but it looks like it may have been a violent tornado. The tornado appears to have tracked across the southern portion of the city and quite wide. It only happened a few hours ago and there are already reports of 24 deaths, but that number will probably rise.

Joplin has a population of about 50,000. This is the latest example of a major tornado striking a densely populated area, an all too frequent occurrence this year.

We have a member from Joplin by the name of JoMo. He was posting in the severe weather thread until the tornado approached. It is unclear whether he was hit but we have not heard from him since the tornado struck. Please keep him in your thoughts.

KNEN - Jacksonville/Whitehouse NOLF FL
KNRA - Coupeville NOLF WA
KO26 - Lone Pine CA
KPHG - Phillipsburg KS
KU14 - Nephi UT
KVMR - Vermillion SD

We may be getting a new Nino in 2025, however whether it's Modoki or canonical is yet to be known. The map currently shows a Modoki.